A while back I ran across a problem 485 communication between a test stand and a product. I had intermittent comm loss during a 40 hour environmental test. I checked all of the system grounds checked my 485 lines swapped out the product and the test equipment still had intermittent loss of communication perhaps once every six hours but never consistent.Then I checked the shield of the 485 cable it looked intact but I noted high resistance between the shield ground and product ground. The shield was terminated to the connector shell. It turned out the connector shell was anodized. The coating used on that connector was non-conductive. A different connector was installed and the communication is now back 100% of the time.
Scoping the analog signals might have detected the noise margin issue faster.
I had a similar disaster when I designed a SCADA system for Bristol Aerospace in Winnipeg MB , Canada in 1977 using the famous Hewlett Packard desktop calculator HP9825. It was for the NRC Rocket Research Range in Churchill and other remote sites. I had two computers connected end-to-end over a 9600 twin pair (RS422) with a limited distance data set. It was a long cable from the Control to Launch Building but I never got to test out my remote control and data collection system live before I finished the project and moved on
It was pretty slick but simple SCADA system and the 1st of its kind in the world. I programmed using the QWERTY keyboard as a live toggle switch for every relay to control power and displayed all the data on two screens of a "dumb terminal" in forms mode with programmable alerts with data flashing in inverse video like a cell in a spreadsheet. (remember this was 1977) It could remote control each experiment and battery pack on the rocket while it was sitting on the launch pad. Up to 10 remote PSU's in the Launch building could be monitored and adjusted in Real Time with the Calc keypad area using +/- to adjust the voltage for umbilical line loss, after selecting the # of the PSU channel. Immediate feedback of the result showed up on the 128 Column LED display and also the slaved Dumb Terminal like a spreadsheet using Forms mode . ADC and DACs were in remote Acquisition cards from HP and some controlled my custom made 96x 25Amp relay box to control power on the dual 50 pin umbilical cables. It also had a high speed magnetic cartridge to save all the current adjustments incase of power failure.
The oversight on my design was the RFI from the Launch control tower causing data errors in the wired OEM communication link. My simple parity checks were inadequate for the error detection task running at 9600 baud full duplex over a 1/4 mile.
I heard later after I left, it did all kinds of weird things on first deployment up North, when the Launch Control guys broadcast on an RF channel.
I don't know what they did with it, but all it needed was Ferrite filters inline to suppress the RF noise on the wired link.
Murphy Strikes again.
I will submit another story on how I resolved the intermitted relay issues on this project later that cause erratic failures. These were slave contacts used to sense the power relay position and man they were flakey. ( but reputable quality parts.)
Anthony Stewart EE'75
Black Brandt Rockets were used to analyze the Ionosphere and Troposphere for NASA prior to the Moon Launch for communication blackouts. a BB-IV can exit the launch tower at 250mph and reach Mach 7 speeds in 30 seconds then float to > 500 miles above the earth before returning to a pre-determined landing area. Black Brandts were only used for data collection in Astro-physics and are still used today. (by Magellan .. nee Bristol Aerospace) . those were exciting days in R&D.
Great find! An excellent example of the 'ah,ha!' moment when the various bits of information come together and success is had. Sounds very familiar to me but I have to regretfully report I didn't often have the success part!
Wow, Pretty good finding that one. Seems like I hear a lot about grounding issues. I wonder if this is a class of issues that is common in distributed systems or systems that involve wider areas? It seems grounding issues might not be so common in components that are smaller and more self contained compared to larger systems that cover more area.
I love this story. Simple dew. The last thing anyone would suspect. Technology interacts with the natural world. In this case it was the moisture in the early morning hours. He noticed while looking at his shoe. Perfect.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.